Medical Discussions

by Drs. Saul Blecker and Raj Khandwalla To read the actual article, click here  Background

• Pts with Hep C have a two percent annual risk of developing hepatocellular carcinoma (HCC)
• HCC detection has a median survival of 4 to 20 months
• No randomized control trials have estabilshed of screening of HCV patients
• Guidelines
  • National Cancer Institute: Fair evidence for no reduction in mortality, good evidence for uncommon but serious harms.
  • New York Department of Health:  Patient with cirrhosis should be screened for hepatocellular carcinoma (HCC) with alpha-fetoprotein (AFP) testing and hepatic ultrasound imaging at least once yearly.

Methods

• Medline search from 1996 to 2002
• Inclusion criteria – randomized control studies and case control studies that determine the sensivity and specificity
• Gold Standard was defined CT, MRI, disease-free greater than 2 years, and/or histopathology. 
• Each study that was included in this review was graded by the three authors

Results

• A total five studies were included.
• A cutoff value of ABP was considered to be 20 mcg/L
• Sensitivity at this level ranged from 41 to 65 percent
• Specificity ranged from 80 to 94 percent
• Positive likelihood ratio was 3.1 to 6.8
• Negative likelihood ratio was 0.4 to 0.6
• Using alternative cutoff ratios
  • Greater than 200 mcg/L –> Sensitivity decreases to 20 to 45 percent and Specificity increases to 99 to 100 percent

Discussion

• Evidence of using AFP is limited and the studies used in this review had questionable methodology
• Given the significant concerns regarding the studies used the authors of the studies could not pool the data to calculate sensitivity and specificity
• The prevalence of HCC in HCV positive patients is 5 percent.
  • Thus, using a cutoff of 20 mcg/L a positive test means that the patients has a 14 to 25 percent probability of having HCC.  A negative test means that the patient has a 2 to 3 percent chance of having HCC.
• If the patient is higher risk (ie known hepatic nodule or liver failure) using a cutoff >200 mcg/L would be very specific for HCC, but at a low level would not rule out HCC.

THE BOTTOM LINE:  Using a cutoff of 20 mcg/L, the sensitivity of measuring alpha-fetoprotein levels in the screening of hepatocellular carcinoma in patients with hepatitis C infection ranges from 41 to 65 percent and the specificity is 80 to 94 percent.  Given the range of these value, widespread screening is debatable.